This is to our knowledge, the largest RCT of IPC ever. It will include more patients and outcome events (proximal DVTs) than all previous RCTs of IPC combined. The patients are enrolled by over 80 hospitals which hopefully will mean that our results will have good external validity. Central randomisation, mainly blinded assessment of our primary outcome, low losses to follow up and intention to treat analysis will minimise bias.
In this trial the primary outcome is an intermediate level outcome. Ideally, we would demonstrate the effect of IPC on symptomatic DVT, PE, survival and functional status. However, symptomatic events are far rarer than those identified by screening CDU and the impact on survival and functional status relatively small. Therefore, a trial would need to enrol many tens of thousands of patients to demonstrate moderate-sized, but clinically important, effects on these outcomes. We can justify the use of an intermediate outcome because of the clinical significance attributed to even asymptomatic DVT in popliteal or femoral veins which most clinicians treat with anticoagulant therapy. Also, IPC is very unlikely to have adverse effects beyond those which will be directly measured in the trial i.e. damage to skin on legs, or falls. IPC should not restrict mobilisation because it can be worn during bed to chair transfers and will be taken off when the patient is able to mobilise independently. Theoretically, IPC might influence blood pressure in the acute phase, which could in turn influence stroke outcome, by increasing venous return - however empirical studies have not demonstrated a significant effect of IPC on blood pressure. The CLOTS trial may underestimate the clinical importance of the effects of IPC because we do not systematically screen for pulmonary emboli with routine imaging. Also, because we systematically screen for asymptomatic DVT, many patients receive treatment with anticoagulants to lessen the risk of symptomatic events occurring.
In stroke patients, and those with other acute medical conditions, IPC can only be applied after the patient has become immobile. Immobility may then persist for weeks or even be permanent in such patients. DVTs may develop rapidly and cannot then be so effectively prevented by a treatment starting sometime after the initial period of paralysis and immobility. Clearly, we cannot test the effectiveness of IPC applied before stroke onset but it is a challenge to recruit patients into a RCT, with the need to collect informed consent, as early after the stroke as IPC might be applied if it was being used in routine clinical practice. For this reason we are trying to maximise the proportion of patients recruited on Day 0 or 1.
Another challenge in this trial is to optimise adherence to the IPC. To be a fair test of the device we need to achieve levels of adherence close to those which would be achieved if we knew that it was effective. However, in a randomised trial one has to achieve a balance between cajoling patients to persist with treatment which they might find uncomfortable whilst allowing them to stop the treatment without having to give a reason as stated in the consent procedure. Inevitably, adherence wanes over time because some patients find the IPC uncomfortable or staff became concerned by the condition of patient's skin. It is for this reason that we are planning to perform a secondary analysis of proximal DVTs up to 14 days.
Symptomatic DVTs affecting the popliteal or femoral veins, which occur within 30 days of randomisation, will be counted in the primary endpoint. A symptomatic DVT is defined for the purposes of this trial as a DVT confirmed on investigation with associated clinical features including leg swelling, pain, obvious erythema or a proven pulmonary embolus. Sometimes these features may not be recognised prior to a positive screening CDU but if present the DVT should be reported as symptomatic. We will distinguish those DVTs which were identified primarily on a screening CDU from those which were diagnosed clinically and confirmed on subsequent investigation since the detection of the latter symptomatic DVTs is not blinded and hence prone to ascertainment bias. A secondary analysis excluding these symptomatic DVTs identified before any CDU will also be performed. Inevitably, some patients will not survive to have routine CDUs and many of these will not have a detailed autopsy to establish whether they had a DVT or PE prior to death. However, it is possible that there will be an imbalance in the number of such deaths between the treatment groups especially if IPC is very effective in reducing the risk of fatal PE. Therefore, we will firstly present the numbers Alive with no DVT, Alive with DVT, Dead without prior DVT, and Missing. We will carry out our analyses in two ways: comparing DVT with no DVT with dead and missing patients excluded; and comparing DVT + dead with no DVT + missing.
We have tried to maximise the benefits to patients from participation in the trial. Our patient information leaflet provides important general information about stroke and its treatment. Our follow up with screening CDU allows us to detect asymptomatic DVTs which may, if undetected and untreated, lead to fatal pulmonary embolism. Lastly we feedback the information gleaned for the 6 month follow-up regarding patients' functional status, mood and pain back to the patients' general practitioners so that this information can inform their future management.
Limitations of the CLOTS trial 3 include the imperfect blinding of the ultrasonographers which could bias detection of our primary outcome, and our inability to blind patients and caregivers which might bias assessment of some of the secondary outcomes. Other limitations included: lack of screening logs; lack of central verification of negative scans or 100% source data verification, but we consider these are unlikely to introduce bias or alter the external validity of the results.